A computing resource may include, for example, a server, a mainframe, a desktop PC, an active cluster, etc. Multiple computing resources may be grouped together to form what is known as a resource pool. A resource pool may be required to execute a variety of requests including, for example, a submitted computing job. A request may originate from any number of sources including a system user, a computing resource outside the resource pool, and/or another resource pool. The resource pool may be expected to execute a request within a certain time frame. The resource pool may also be expected to be available for executing requests at specific times. Expectations associated with a resource pool may be set forth in a service level agreement (SLA).
The number of requests requiring execution may be considered, at least in part, the computational load associated with the resource pool. A computational load may vary with time. For example, a computational load may reach a peak during certain times of the day, but may be almost zero at other times. The expectations set forth in a SLA may need to be satisfied at all times and regardless of the computational load.
A resource pool may require electrical power to execute requests. The electrical power consumed may be dependent on the number of resources in the resource pool actively executing requests. The consumption of electrical power may result in the generation of heat. Accordingly, a resource pool executing requests may need to be cooled by a cooling system. However, a cooling system may also consume electrical power while attempting to reduce the heat generated by the resource pool.